Notes on Poisson distribution-based minimum error thresholding

被引：14
作者：
Fan, JL [1 ]
机构：
[1] Xian Inst Posts & Telecommun, Dept Basic Courses, Xian 710061, Peoples R China
来源：
PATTERN RECOGNITION LETTERS | 1998年 / 19卷 / 5-6期
关键词：
Poisson distribution; minimum error thresholding; relative entropy; maximum likelihood principle;
D O I：
10.1016/S0167-8655(98)00028-2
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Pal and Bhandari [On object background classification. Int. J. Syst. Sci. 23 (1992) 1903-1920] modified the minimum error threshold originated by Kittler and Illingworth [Minimum error thresholding. Pattern Recognition 19 (1986) 41-47], proposing the Poisson distribution-based minimum error threshold. In this paper, Poisson distribution-based minimum error thresholding is explained using relative entropy and the maximum likelihood principle. (C) 1998 Elsevier Science B.V. All rights reserved.
引用
收藏
页码：425 / 431
页数：7
相关论文
共 12 条
[1]   Minimum cross-entropy threshold selection [J].
Brink, AD ;
Pendock, NE .
PATTERN RECOGNITION, 1996, 29 (01) :179-188
[2]   A RELATIVE ENTROPY-BASED APPROACH TO IMAGE THRESHOLDING [J].
CHANG, CI ;
CHEN, K ;
WANG, JW ;
ALTHOUSE, MLG .
PATTERN RECOGNITION, 1994, 27 (09) :1275-1289
[3]  
Dainty JC., 1974, IMAGE SCI
[4]   AXIOMATIC CHARACTERIZATION OF THE DIRECTED DIVERGENCES AND THEIR LINEAR COMBINATIONS [J].
JOHNSON, RW .
IEEE TRANSACTIONS ON INFORMATION THEORY, 1979, 25 (06) :709-716
[5]   MINIMUM ERROR THRESHOLDING [J].
KITTLER, J ;
ILLINGWORTH, J .
PATTERN RECOGNITION, 1986, 19 (01) :41-47
[6]  
KULLBACK S, 1968, INFORMATION THEORY S
[7]   MAXIMUM-LIKELIHOOD THRESHOLDING BASED ON POPULATION MIXTURE-MODELS [J].
KURITA, T ;
OTSU, N ;
ABDELMALEK, N .
PATTERN RECOGNITION, 1992, 25 (10) :1231-1240
[8]  
LE CH, 1993, PATTERN RECOGN, V26, P617
[9]   THRESHOLD SELECTION METHOD FROM GRAY-LEVEL HISTOGRAMS [J].
OTSU, N .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS, 1979, 9 (01) :62-66
[10]  
Pal N. R., 1991, INT J PATTERN RECOGN, V5, P439
12